Are We Overconfident in Models and Results for Semi-Supervised 3D Medical Image Segmentation?
For researchers in medical image segmentation, this work highlights and addresses critical methodological flaws that may inflate reported performance, advocating for more rigorous evaluation protocols.
The paper identifies a twofold overconfidence problem in semi-supervised 3D medical image segmentation: algorithmic confirmation bias from conflating confidence with uncertainty, and strategic overfitting due to using test sets for validation. The proposed TCSeg framework decouples confidence from uncertainty and corrects bias across three spaces, achieving strong performance on three benchmarks.
Semi-supervised learning has become a dominant paradigm for reducing annotation costs. However, we argue that the current progress is clouded by a twofold overconfidence problem. Algorithmically, mainstream pseudo-labeling frameworks often conflate prediction confidence with uncertainty, leading to severe confirmation bias. Strategically, since multiple benchmark datasets lack dedicated validation sets, some studies use the test set for validation as well, leading to inflated performance estimates. Subsequent methods, compelled to employ the same strategy to surpass reported SOTA, trigger an arms race of overfitting. This raises concerns that the impressive numerical gains in the community may reflect overfitting rather than genuine progress. Thus, we propose a tri-space calibrated segmentation framework founded on a principled dual-axis reliability assessment engine. It explicitly decouples confidence from uncertainty and uses this signal to detect and correct confirmation bias across feature, probability, and image spaces in a collaborative manner. Across three benchmark datasets, TCSeg consistently delivers strong performance under existing evaluation protocols. More importantly, we advocate that the community report final-checkpoint results under multiple-run protocols, thereby establishing more rigorous benchmarks with a more realistic perspective. Code will be available: github.com/DirkLiii/TCSeg.